Download Inability of the Polyphasic Approach to Systematics To Determine

INTERNATIONAL
JOURNAL
OF SYSTEMATIC
BACTERIOLOGY,
Apr. 1995, p. 379-381
0020-7713/95/$04.00+0
Copyright 0 1995, International Union of Microbiological Societies
Vol. 45, No. 2
Inability of the Polyphasic Approach to Systematics To
Determine the Relatedness of the Genera
Xenorhabdus and Photorhabdus
F. A. RAINEY,l R.-U. EHLERS,2 AND E. STACKEBRANDT'"
DSM-Deutsche Sammlung von Mikroorganismen und Zellkulturen GmbH, 38124 Braunschweig,
and Abteilung Biotechnologie und Biologischer Pjlanzenschutz, Institut fur Phytopathologie,
Christian-Albrechts-Universitat, 24223 Klausdoi$ Germany
Comparative analysis of the genes coding for 16s rRNA of the type strains of Xenorhubdus and Photorhubdus
species indicates the close phylogenetic relationship of these two genera. However, distance matrix analyses do
not unambiguously separate the symbionts of entomopathogenic nematodes according to their assignment into
different genera. When various 16s rRNA gene sequences from a selection of members of the gamma subclass
of Proteobucteriu and outgroup taxa were used, the intrageneric relationships of Xenorhubdus species and the
positions of Photorhubdus luminescens and related species changed significantly.
The genus Photorhabdus has recently been described for a
group of bioluminescent and facultatively anaerobic insectpathogenic strains that are symbionts of entomopathogenic
nematodes belonging to the family Heterorhabditidae (3). The
exclusion of Photorhabdus luminescens from the genus Xenorhabdus was based on low interspecies DNA reassociation values
(3, 6, 7, 16) and the presence of distinguishing phenotypic (2,
3) and chemotaxonomic (16) properties. Thus, the genus Xenorhabdus was emended to contain only strains isolated from
different entomopathogenic nematode species of the family
Heterorhabditidae. None of the studies, however, were able to
determine the intergeneric relatedness of these two taxa. Similarly, 16s rRNA cataloguing performed on two strains of P.
(Xenorhabdus)luminescens and one X nematophilus strain (5)
was carried out not to determine the possible phylogenetic
heterogeneity of the genus Xenorhabdus, but to investigate
whether these organisms belong to the family Enterobacteriaceae. In this study, we determined the 16s rRNA gene (16s
rDNA) sequences of all type strains of both genera, and we
discuss the problem of unambiguous phylogenetic placement
of organisms.
reconstructed by applying additive treeing methods, such as neighbor joining (12)
and the algorithm of De Soete (4), using the corrected dissimilarity values. One
thousand bootstrap values for 920 polymorphic sites were calculated to test the
stability of the neighbor-joining tree.
Nucleotide sequence accession numbers. The 16s rDNA sequences have been
deposited at EMBL under accession no. X82248 to X82254.
RESULTS
The almost complete sequences of the seven strains of Xenorhabdus and Photorhabdus species were analyzed. Binary
16s rDNA similarities for the Xenorhabdus strains ranged between 96.0 and 97.7%, while the values for these organisms
and P. luminescens ranged between 94.1 and 96.6%. With 92.3
to 95.2% 16s rDNA similarity, the membership of Xenorhabdus and Photorhabdus in the family Enterobacteriaceae is in
accord with previous findings (5). Proteus vulgaris can be considered the nearest phylogenetic neighbor (between 93.5 and
95.1% similarity). Other members of this subclass, including
Vibrio parahaemolyticus, selected as an out-group member of
the gamma subclass of Proteobacteria, have significantly less
similarity, mostly <90% (data not shown).
Independently of the number of generic affiliation of outgroup reference organisms, two of the most widely used algorithms for generating phylogenetic trees from dissimilarity values, i.e., the distance matrix methods of Saitou and Nei (12)
and De Soete (4), consistently placed strains of Xenorhabdus
and Photorhabdus on a single line of descent, separate from
other members of the family Enterobacteriaceae. The clustering is supported by 100% significance as determined by bootstrap analysis. The number of signature nucleotides in the 16s
rDNAs of Photorhabdus and Xenorhabdus strains is small, i.e.,
at positions 19 to 916 (C-U versus A-U, found in other enterobacteria) and positions 154 to 167 (C-G versus U-A, found in
most other enterobacteria). A few signature nucleotides are
shared with Proteus vulgaris, such as those at positions 320 to
333 (G-C versus A-U), 576 to 647 (A-U versus U-A), 577 to
646 (A-U versus G-C), and 682 to 707 (Y-G versus G-C). The
G + C contents of 16s rDNAs of both genera of the symbionts
of entomopathogenic nematodes are about 2 to 3% higher
(55.1 to 55.5%) than those of other members of the Enterobacteriaceae (52.8 to 53.4%).
With a selection of nine species of the family Enterobacteriaceae and I/: parahaemolyticus used as reference organisms, a
neighbor-joining tree was generated (Fig. 1). In contrast to the
MATERIALS AND METHODS
Strains investigated. Xenorhabdus nematophilus DSM 3370T, X. beddingii
DSM 4764T,X. poinarii DSM 4768T, X. bovienii 4766=, and Photorhabdus luminescens DSM 3368T and DSM 3369 are deposited in the DSM-German Collection of Microorganisms and Cell Cultures. P. luminescens HSH2 was isolated by
R.-U. Ehlers from the nematode Heterorhabditis megidis HSH2. All strains were
cultivated on medium 423 as indicated in the DSM catalogue of strains.
Analysis of 16s rDNk Extraction of genomic DNA and amplification of the
16s rDNA were carried out as described previously (11). PCR products were
sequenced directly by using the Tuq DyeDeoxy Terminator Cycle sequencing kit
(Applied Biosystems) according to the manufacturer's protocol. The sequence
reaction mixtures were electrophoresed by using the Applied Biosystems 373A
DNA sequencer.
Phylogenetic analysis. 16s rDNA sequences were compared with the existing
16s rDNA database of members of the family Enterobucteriaceae and other
members of the gamma subclass of Proteobacteria (9). Dissimilarity values were
transformed into phylogenetic distance values that compensate for multiple
substitutions at any given site in the sequence (8). Phylogenetic trees were
* Corresponding author. Mailing address: DSM-Deutsche Sammlung von Mikroorganismen und Zellkulturen GmbH, Mascheroder
Weg lb, 38124 Braunschweig, Germany. Phone: 49 531 2616 352.
Fax: 531 2616 418. Electronic mail address: [email protected].
379
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Mon, 19 Jun 2017 02:21:40
380
INT.J. SYST.BACTERIOL.
RAINEY ET AL.
Xenorhabdus bovienii
-I
Xenorhabduspoinarii
::
;. 37
3
7
-
Xenorhabdus beddingii
Xenorhabdus nematophilus
-Photorhabdus luminescens strain HSH2
Photorhabdus luminescens DSM 33681"
Photorhabdus luminescens DSM 3369
,
1 4-
l l
Proteus vulgaris
Hafniaalvei
I
Yersinia enterocolitica
I
Pantoea agglornerans
Erwinia carotovora
Serratia marcescens
I
Citrobacterfreundii
Escherichia coli
Plesiomonas shigelloides
FIG. 1. Neighbor-joining tree showing the phylogenetic relationships among members of the genera Xenorhabdus and Photorhabdus and members of the family
Enterobacteriaceae. 1.: parahaemolyticus served as a root. The region in which the order of branching points leading to species of Xenorhabdus and Photorhabdus cannot
be determined with confidence is indicated (shaded box). The bar represents five nucleotide substitutions per 100 nucleotides.
distinctly lower 16s rDNA similarity of about 2%, the three
strains of P. luminescens do not cluster apart fromxenorhabdus
strains but, rather, branch off from within the radiation of
strains of the latter genus. While X nematophilus and X beddingii form a pair and X. bovienii and X. poinarii form a pair of
distantly related species, supported by bootstrap values of 37
and 93%, respectively, X poinarii and X. bovienii branch more
deeply than the three P. luminescens strains. This branching
pattern is not in accord with the bootstrap analysis, which
indicated that, to the exclusion of P. luminescens, the type
strains of Xenorhabdus clustered together in 65% of simulated
trees. The same topology was obtained when the tree was
generated by using the algorithm of De Soete (4). To investigate whether the number and origin of homologous sequences
from various taxa of the gamma subclass influence the branching point of the Xenorhabdus and Photorhabdus species, 15
different sequence sets were analyzed by the two distance matrix algorithms. The number of sequences ranged from 8 to 40,
and the selection of sequences included those of the symbiotic
bacteria of the genus Buchnera (10) as well as those of the most
deeply branching taxa of this subclass, i.e., sequences of species
of Nitrosomonas, Coxiella,Xanthomonas, Chromatium, and Ectothiorhodospira.With two exceptions, in which the two genera
investigated form phylogenetic sister groups, P. luminescens
grouped within the radiation of Xenorhabdus species. This was
independent of whether only the type strain of P. luminescens
or all three strains of this species were included in the analysis.
In most of these cases, P. luminescens clustered with the pair X.
nematophilus-X. beddingii, and X bovienii clustered most
deeply. This situation is similar to the one depicted in Fig. 1.
The exceptions are those cases in which, in addition to the
sequences of Xenorhabdus and Photorhabdus strains, randomly
selected sequences of either only Escherichia coli or of all
enterobacteria, three sequences of Buchnera strains, and one
representative each of the genera Vibrio and authentic Pseudomonas were used (not shown). In order to demonstrate the
inability of the algorithms and reference sequences used to
reliably depict the relatedness between members of the genera
Xenorhabdus and Photorhabdus, the region of uncertain
branching points is indicated in Fig. 1.
DISCUSSION
Justification for the exclusion of X. luminescens from the
genus Xenorhabdus and the description of Photorhabdus to
accommodate this species was based on results of DNA-DNA
pairing experiments, as well as chemotaxonomic and phenotypic characteristics. The intrageneric relatedness of Xenorhabdus species has been determined previously by different
DNA-DNA reassociation methods (3, 6, 7, 16). In all of these
studies, DNA relatedness determined for the type strains of X
nematophilus and X luminescens was <20%, while the similarity between the type strains of the other Xenorhabdus species
ranged between 20 and 40%. As DNA reassociation values for
use in genus delineation are lacking, and low similarity values
alone do not justify the dissection of a genus, further evidence
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Mon, 19 Jun 2017 02:21:40
VOL.45, 1995
RELATEDNESS OF THE GENERA XENORHABDUS AND PHOTORHABDUS
for the exclusion of X. luminescens was needed. While the
ubiquinone composition and the DNA G + C content were not
able to discriminate between these two taxa, differences in the
major cellular fatty acids had been reported (16). However, the
strains of X. luminescens differed from the other members of
the genus only in having a slightly higher content of i-15
branched fatty acids. Results of a numerical taxonomic study
showed that the symbionts of Heterorhabditis spp., X luminescens strains, form a discrete cluster regardless of the datum set
and the clustering strategy used. Thus, the phenetic data are in
accordance with the DNA pairing study, the fatty acid analysis,
and the 16s rDNA analysis. Strains of X. luminescens were
linked most closely to the Xenorhabdus strains isolated from
Steinemema organisms of Australian origin (group V as defined by Akhurst and Boemare [ 11) which were described as X.
beddingii. In three of four numerical analyses X. luminescens
and X. beddingii formed a sister taxon to strains of X. bovienii,
and in only one analysis strains of X. luminescens branched
earlier than strains of the other species of Xenorhabdus. It thus
appears that the exclusion of X. luminescens from the genus
Xenorhabdus and the description of P. luminescens were done
mainly on the basis of some phenotypic data in which strains of
this species differed exclusively from strains of Xenorhabdus,
such as presence of bioluminescence, positive catalase reaction, assimilation of arbutin, lack of DL-lactate production, and
the host nematode. However, not all strains of P. luminescens
are luminescent (l), and some strains closely related to X
bovienii are also catalase positive.
We do not want to question the validity of the transfer of X.
luminescens to a new genus. The 16s rDNA analysis reflects
the situation of the numerical phenetic analysis closely in that
the relationship of P. luminescens to the species of Xenorhabdus depends heavily on the database used. Thus, if one agrees
to this taxonomic conclusion, the two genera can nevertheless
be considered very closely related, to the extent that the degree
of genomic differences observed between the type species of
the two genera is not significantly higher than that found between the species of Xenorhabdus.
The finding that the selection of reference organisms influences tree topologies quite significantly has been observed
previously (14, 17). In general, every new sequence which is
different from those in the existing datum set provides additional data on the information content of individual sequence
positions as well as on the characteristics defining phylogenetic
groups. The inclusion of this additional information may
change the tree locally or even globally. This effect is recognized by those who construct phylogenetic trees from sequence
data. Of the most obvious parameters, we can exclude a bias
introduced by differences in the base composition of rDNA, as
all Xenorhabdus and Photorhabdus species have very similar
G+C values. Bias introduced by partial sequences (13) can be
excluded, as almost complete sequences have been compared.
However, we have noted that Photorhabdus species may be
subjected to a slightly higher evolutionary rate (“fast clock” or
tachytelic behavior). Support for this hypothesis originates
from the finding that the level of similarity between these
strains and the reference organism is about 2% lower than that
found for Xenorhabdus species and the reference organisms
(which should be very similar in comparison with out-group
sequences) (15). At the nucleotide level, this hypothesis is
supported by the lack of several absolutely unique positions in
381
the sequences of P. luminescens which are conserved for members of the gamma subclass of Proteobacteria (nucleotides 1001
to 1002 pairing with nucleotides 1038 to 1039 and nucleotides
1006 to 1009 pairing with nucleotides 100 to 1023; bp 1254 to
1283). Furthermore, strains of P. luminescens show greater
variation in the nucleotide composition of hypervariable regions than is found among the type strains of the Xenorhabdus
species.
The strategy that should be considered next is to increase the
stability of the tree by adding more sequences of members of
the two species, especially more strains of P.luminescens. If, as
expected, P. luminescens strains are slightly more rapidly evolving, then this should be reflected and detectable in a greater
variation of those nucleotide positions considered to be conserved in neighboring taxa.
REFERENCES
1. Akhurst, R. J., and N. E. Boemare. 1988. A numerical taxonomic study of the
genus Xenorhabdus (Enterobacteriaceae) and proposed elevation of the subspecies of X. nematophilus to species. J. Gen. Microbiol. 1341835-1845.
2. Boemare, N. E., and R. J. Akhurst. 1988. Biochemical and physiological
characterization of colony form variants in Xenorhabdus spp. (Enterobacteriaceae). J. Gen. Microbiol. 134751-761.
3. Boemare, N. E., R. J. Akhurst, and R. G. Mourant. 1993. DNA relatedness
between Xenorhabdus spp. (Enterobacteriaceae),symbiotic bacteria of entomopathogenic nematodes, and a proposal to transfer X . luminescens to a new
genus, Photorhabdus gen. nov. Int. J. Syst. Bacteriol. 43:249-255.
4. De Soete, G. 1983. A least squares algorithm for fitting additive trees to
proximity data. Psychometrika 48621-626.
5. Ehlers, U.-R., U. Wyss, and E. Stackebrandt. 1988. 16s rRNA cataloging and
the phylogenetic position ofXenorhabdus. Syst. Appl. Microbiol. 10:121-125.
6. Farmer, J. J., J. H. Jorgensen, P. A. Grimont, R. J. Akhurst, G. 0. Poinar,
E. Ageron, G. V. Pierce, J. A. Smith, G. P. Carter, K. L. Wilson, and F. W.
Hickman-Brenner. 1989. Xenorhabdus luminescens (DNA hybridization
group) from human clinical specimens. J. Clin. Microbiol. 271594-1602.
7. Grimont, P. A, A. G. Steigemalt, N. E. Boemare, F. W. Hickman-Brenner,
C. Deval, F. Grimont, and D. Brenner. 1984. Deoxyribonucleic acid relatedness and phenotypic study of the genus Xenorhabdus. Int. J. Syst. Bacteriol.
34378-388.
8. Jukes, T. H., and C. R. Cantor. 1969. Evolution of protein molecules, p.
21-132. In H. N. Munro (ed.), Mammalian protein metabolism. Academic
Press, New York.
9. Larsen, N., G. J. Olsen, B. L. Maidak, M. McCaughey, J. R. Overbeek, T. J.
Macke, T. L. Marsh, and C. R Woese. 1993. The ribosomal database project.
Nucleic Acids Res. 21:3021-3023.
10. Munson, M. A., P. Baumann, and M. G. Kinsey. 1991. Buchneru gen. nov.
and Buchnera uphidicola sp. nov., a taxon consisting of the mycetocyteassociated, primary endosymbionts of aphids. Int. J. Syst. Bacteriol. 41566568.
11. Rainey, F. A,, M. Dorsch, H. W. Morgan, and E. Stackebrandt. 1992. 16s
rDNA analysis of Spirochaeta themophila: its phylogenetic position and
implications for the systematics of the order Spirochaetales. Syst. Appl. Microbiol. 15197-202.
12. Saitou, N., and M. Nei. 1987. The neighbor-joining method: a new method
for reconstructing phylogenetic trees. Mol. Biol. Evol. 4406-425.
13. Stackebrandt, E., W. Liesack, and D. Witt. 1992. Ribosomal RNA and
ribosomal DNA sequence analyses. Gene 115255-260.
14. Stackebrandt, E., and W. Ludwig. 1994. The importance of using outgroup
reference organisms in phylogenetic studies: the Atopobium case. Syst. Appl.
Microbiol. 1739-43.
15. Stackebrandt, E., and C. R. Woese. 1981. The evolution of prokaryotes, p.
1-31. In M. J. Carlile, J. F. Collins, and B. E. B. Moseley (ed.), Molecular
and cellular aspects of microbial evolution. Cambridge University Press,
Cambridge.
16. Suzuki, T., S. Yamanaka, and Y. Nishimura. 1990. Chemotaxonomic study
of Xenorhabdus species+ellular fatty acids, ubiquinone and DNA-DNA
hybridization. J. Gen. Appl. Microbiol. 36393-401.
17. Woese, C. R., L. Achenbach, P. Rouviere, and L. Mandelco. 1991. Archaeal
phylogeny: reexamination of the phylogenetic position of Archueogfobus
filgidus in the light of certain composition-induced artifacts. Syst. Appl.
Microbiol. 14364-371.
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Mon, 19 Jun 2017 02:21:40